Oxford researchers study AI to help them diagnose prostate cancer

Researchers and clinicians at the University of Oxford have begun an evaluation of AI software that could help pathologists diagnose prostate cancer. The UK reports around 46,000 new prostate cancer cases yearly, a 12 percent increase in the past 10 years.

Testing of the technology in a clinical setting, the university announced in a press release, is underway at Oxford University Hospitals (OUH) NHS Foundation Trust. It represents a key milestone in the University of Oxford-led ARTICULATE PRO study.

This two-year project aims to investigate the deployment of AI in the prostate cancer pathway by using Paige Prostate. The tool is a computer-assisted diagnostic system that seeks to help pathologists detect, grade, and measure tumors in prostate biopsies or tissue samples. Two other NHS trusts – University Hospital Coventry and Warwickshire and North Bristol NHS Trust – will also assess the Paige Prostate Suite software.

Portuguese scientists develop a new optical technique for early cancer diagnosis

Researchers from the University of Coimbra, Portugal, announced a new optical technique that helps identify cancer in patients in their early stages. This method can utilize laser radiation or infrared to detect cancer and even analyze the surgical margin of each case.

Aiming for accurate diagnosis

The technology should immediately flag suspicious areas to pathologists by identifying the hallmarks of malignant cells captured by previous training in large datasets of biopsies. It also assesses the amount of tumor present and how aggressive it appears.

The project is led by Oxford Associate Professor Clare Verrill, an expert urological histopathologist. She said: “One of our key aims in the health service is to diagnose cancers accurately and at an earlier stage so that treatment can be delivered more quickly and, ultimately, patient outcomes improve. It will be great news for patients if we can harness this diagnostic technology to achieve this”.

This algorithm can predict prostate cancer risk

Cambridge scientists have developed a tool for predicting a person’s risk of developing prostate cancer. The tool is called CanRisk-Prostate and was developed by researchers at the University of Cambridge and The Institute of Cancer Research.

Making clinicians and patients accustomed

“That’s why this evaluation – one of the first of its kind – is an important step. We will be looking not only at how well this software performs in a busy clinical setting and whether diagnostic accuracy and efficiency improves but also assessing the experience of clinicians and patients and the impact on workflow”. This way, OUH pathologists use AI applications to help read prostate biopsy slides as part of their routine work.

She added: “In 2020, OUH’s histopathology laboratory was one of the first in the UK to achieve the milestone of scanning 100 percent of its surgical histology workload. Our digital pathology experience makes us an ideal setting to test AI technologies such as Paige Prostate Suite in a real-world clinical setting”.